New energy commercial vehicle dynamic path optimization method and system based on AI

By constructing a distance, energy consumption, and time matrix for new energy commercial vehicles and combining it with a genetic algorithm to optimize route planning, the multi-objective optimization problem of new energy commercial vehicles under dynamic customer nodes was solved, improving operational efficiency and customer satisfaction.

CN120893645APending Publication Date: 2025-11-04SHENZHEN QIXING FENGTAI TECHNOLOGY CO LTD

Patent Information

Application Number
CN202511024292.7
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2025-07-24
Publication Date
2025-11-04

AI Technical Summary

Technical Problem

Existing path optimization methods focus on a single objective and are difficult to adapt to the dynamic changes in customer nodes of new energy commercial vehicles, resulting in low operating efficiency.

Method used

By obtaining the route map of the target area, marking the effective node set, constructing distance, energy consumption, and time matrices, optimizing route planning using expression trees and genetic algorithms, and dynamically adjusting it in conjunction with a customer importance list.

Benefits of technology

It achieves multi-objective optimization, improves the operating efficiency of new energy commercial vehicles, ensures the priority of meeting the needs of important customers, reduces ineffective mileage and energy consumption costs, and improves operational stability and efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN120893645A_ABST
    Figure CN120893645A_ABST
Patent Text Reader

Abstract

The invention discloses an AI-based new energy commercial vehicle dynamic path optimization method and system, and relates to the technical field of path planning. Determining a driving path between every two effective nodes in the route map to obtain a distance matrix; obtaining an energy consumption matrix according to the distance matrix and the energy consumption rate of the new energy commercial vehicle, and obtaining a time matrix according to the distance matrix and the driving speed of the new energy commercial vehicle; initializing an effective node set and a function node set, and randomly combining effective nodes and function nodes to construct a plurality of expression trees; sorting all client nodes according to the distance matrix, the energy consumption matrix and the time matrix to obtain a client importance list, and substituting the client importance list into a genetic algorithm to determine a route plan of each new energy commercial vehicle; after the new energy commercial vehicles drive one node, the route of each new energy commercial vehicle is dynamically planned, and the operation efficiency of the new energy commercial vehicles is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention belongs to the field of path planning technology, specifically relating to an AI-based dynamic path optimization method and system for new energy commercial vehicles. Background Technology

[0002] With the continuous development of new energy vehicle technology, new energy commercial vehicles are increasingly widely used in logistics and transportation. However, route planning for new energy commercial vehicles faces numerous challenges. Compared with traditional fuel vehicles, they are more affected by factors such as energy consumption and charging. Furthermore, customer nodes often change dynamically during actual transportation, significantly increasing the difficulty of route planning. Currently, most existing route optimization methods are designed for static scenarios and are ill-suited to adapting to dynamic changes in customer nodes. Moreover, when considering optimization objectives, they often focus only on a single objective, such as time or cost, failing to achieve coordinated optimization of multiple objectives such as energy consumption, time, and cost. In addition, there are shortcomings in addressing the unique charging constraints of new energy commercial vehicles and the reasonable consideration of customer priorities, resulting in poor efficiency and economy in route planning.

[0003] Patent CN118129789A discloses a path planning method for new energy vehicles, including the following steps: S1, obtaining the current driving range of the new energy vehicle, and determining a charging station based on the vehicle's starting point, destination, and driving range; S2, generating the first segment of the driving path between the vehicle's starting point and the charging station; S3, constructing a fuzzy matrix to generate the final segment of the driving path between the charging station and the destination, thus completing the path planning. However, this solution still focuses only on a single objective and is difficult to adapt to dynamic changes in customer nodes, resulting in low operating efficiency for new energy commercial vehicles. Summary of the Invention

[0004] The purpose of this invention is to address the problem of low operating efficiency of new energy commercial vehicles due to focusing on a single objective and being unable to adapt to dynamic changes in customer nodes. Therefore, this invention proposes an AI-based dynamic path optimization method and system for new energy commercial vehicles.

[0005] In a first aspect of this invention, an AI-based dynamic path optimization method for new energy commercial vehicles is proposed, the method comprising: Obtain a route map of the target area and mark all valid nodes in the route map to obtain a set of valid nodes; the valid nodes include customer nodes, charging station nodes, and warehouse nodes; the warehouse nodes include multiple new energy commercial vehicles; The distance matrix is ​​obtained by determining the driving paths between each pair of valid nodes in the set of valid nodes. An energy consumption matrix is ​​obtained based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and a time matrix is ​​obtained based on the distance matrix and the driving speed of the new energy commercial vehicle. Initialize the set of valid nodes and the set of function nodes, and randomly combine the valid nodes and function nodes to construct multiple expression trees; The fitness of each expression tree is calculated based on the distance matrix, energy consumption matrix, and time matrix. All expression trees are sorted from largest to smallest fitness, and the top preset number of expression trees are obtained for updating and optimization to obtain the optimal expression tree. The customer importance list is obtained by sorting all customer nodes according to the optimal expression tree, and the customer importance list is substituted into the genetic algorithm to determine the route planning for each new energy commercial vehicle. After each new energy commercial vehicle completes a node, the customer node in the route map is updated, and the route of each new energy commercial vehicle is dynamically planned based on the update result.

[0006] Optionally, the energy consumption matrix is ​​obtained based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and the time matrix is ​​obtained based on the distance matrix and the driving speed of the new energy commercial vehicle, including: The energy consumption rate of the new energy commercial vehicle is determined based on historical data, and the energy consumption matrix is ​​obtained by multiplying the energy consumption rate by the distance matrix. The average driving speed of the new energy commercial vehicle is determined based on historical data, and the time matrix is ​​obtained by multiplying the average driving speed by the distance matrix.

[0007] Optionally, calculating the fitness of each expression tree based on the distance matrix, energy consumption matrix, and time matrix includes: Through formula Obtain the fitness F of each expression tree; Where F is the fitness of the expression tree. , and These are weighted coefficients, and their sum is 1. The total energy consumption of the expression tree. The total cost of the expression tree. The total time for the expression tree. Let be the energy consumption from node i to node j, and n be the number of nodes in the path in the expression tree. The variable is a decision variable; it is 1 if the path contains an edge from i to j, and 0 otherwise. Cost per unit of electricity Charging cost coefficient, Let $\mathbf$ be the amount of charge from node $i$ to node $j$. Here, P represents the time window penalty coefficient. Let be the travel time from node i to node j, and m be the number of client nodes in the expression tree. The service time for the kth customer.

[0008] Optionally, sorting all customer nodes according to the optimal expression tree to obtain a customer importance list, and combining this with a genetic algorithm to determine the route planning for each new energy commercial vehicle includes: Multiple initial route schemes are generated based on the customer importance list, resulting in multiple sets of parallel chromosomes; Parallel genetic operations are performed iteratively on multiple sets of parallel chromosomes until the stopping condition is met to obtain the target parallel chromosome. The route planning for each new energy commercial vehicle is determined based on the target parallel chromosome.

[0009] Optionally, after each new energy commercial vehicle completes a node, the customer node in the route map is updated, and dynamic route planning is performed for each new energy commercial vehicle based on the update result, including: The customer nodes in the roadmap are updated to obtain updated customer nodes; the update includes removing customer nodes in the roadmap that have completed their tasks and adding new customer nodes. Based on the updated client node, regenerate the updated optimal expression tree and calculate the similarity between the updated optimal expression tree and the optimal expression tree; If the similarity is greater than or equal to a preset value, the route of each new energy commercial vehicle remains unchanged; If the similarity is less than a preset value, the route of each new energy commercial vehicle is re-determined based on the updated optimal expression tree and the genetic algorithm.

[0010] In a second aspect of this invention, an AI-based dynamic path optimization system for new energy commercial vehicles is proposed, comprising: The effective node set determination module is used to obtain a route map of the target area and mark all effective nodes in the route map to obtain an effective node set; the effective nodes include customer nodes, charging station nodes and warehouse nodes; the warehouse nodes include multiple new energy commercial vehicles; The first matrix determination module is used to determine the driving path between each pair of valid nodes in the set of valid nodes to obtain a distance matrix; The second matrix determination module is used to obtain an energy consumption matrix based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and to obtain a time matrix based on the distance matrix and the driving speed of the new energy commercial vehicle. An expression tree construction module is used to initialize the set of valid nodes and the set of function nodes, and to randomly combine valid nodes and function nodes to construct multiple expression trees; The optimal expression tree determination module is used to calculate the fitness of each expression tree based on the distance matrix, energy consumption matrix, and time matrix, sort all expression trees from largest to smallest fitness, and obtain the top preset number of expression trees for updating and optimization to obtain the optimal expression tree; The route planning module is used to sort all customer nodes according to the optimal expression tree to obtain a customer importance list, and then substitute the customer importance list into the genetic algorithm to determine the route planning for each new energy commercial vehicle. The route update module is used to update the customer nodes in the route map after each new energy commercial vehicle has completed a node, and to dynamically plan the route for each new energy commercial vehicle based on the update results.

[0011] Optionally, the second matrix determination module includes: The energy consumption matrix determination module is used to determine the energy consumption rate of the new energy commercial vehicle based on historical data, and obtain the energy consumption matrix by multiplying the energy consumption rate by the distance matrix. The time matrix determination module is used to determine the average driving speed of the new energy commercial vehicle based on historical data, and to obtain the time matrix by multiplying the average driving speed by the distance matrix.

[0012] Optionally, the optimal expression tree determination module includes: Through formula Obtain the fitness F of each expression tree; Where F is the fitness of the expression tree. , and These are weighted coefficients, and their sum is 1. The total energy consumption of the expression tree. The total cost of the expression tree. The total time for the expression tree. Let be the energy consumption from node i to node j, and n be the number of nodes in the path in the expression tree. The variable is a decision variable; it is 1 if the path contains an edge from i to j, and 0 otherwise. Cost per unit of electricity Charging cost coefficient, Let $\mathbf$ be the amount of charge from node $i$ to node $j$. Here, P represents the time window penalty coefficient. Let be the travel time from node i to node j, and m be the number of client nodes in the expression tree. The service time for the kth customer.

[0013] Optionally, the route planning module includes: A multi-parallel chromosome generation module is used to generate multiple initial route schemes based on the customer importance list to obtain multiple sets of parallel chromosomes. The target parallel chromosome determination module is used to iteratively perform parallel genetic operations on multiple sets of parallel chromosomes until the stopping condition is met to obtain the target parallel chromosome. The new energy commercial vehicle route determination module is used to determine the route plan for each new energy commercial vehicle based on the target parallel chromosome.

[0014] Optionally, the route update module includes: The customer node update module is used to update the customer nodes in the roadmap to obtain updated customer nodes; the update includes removing customer nodes in the roadmap that have completed tasks and adding new customer nodes; The similarity calculation module is used to regenerate the updated optimal expression tree based on the updated client node, and calculate the similarity between the updated optimal expression tree and the optimal expression tree. The route-unchanged module is used to ensure that the route of each new energy commercial vehicle remains unchanged if the similarity is greater than or equal to a preset value. The route change module is used to redetermine the route of each new energy commercial vehicle based on the updated optimal expression tree and the genetic algorithm if the similarity is less than a preset value.

[0015] The beneficial effects of this invention are: This invention proposes an AI-based dynamic path optimization method for new energy commercial vehicles. The method involves obtaining a route map of the target area and marking all valid nodes in the route map to obtain a valid node set. Valid nodes include customer nodes, charging station nodes, and warehouse nodes. Warehouse nodes include multiple new energy commercial vehicles. The method determines the travel paths between each pair of valid nodes in the valid node set to obtain a distance matrix. An energy consumption matrix is ​​obtained based on the distance matrix and the energy consumption rate of the new energy commercial vehicles. A time matrix is ​​obtained based on the distance matrix and the travel speed of the new energy commercial vehicles. The method initializes the valid node set and the function node set, and randomly combines valid nodes and function nodes to construct multiple expression trees. The fitness of each expression tree is calculated based on the distance matrix, energy consumption matrix, and time matrix. All expression trees are sorted from largest to smallest fitness, and the top preset number of expression trees are updated and optimized to obtain the optimal expression tree. All customer nodes are sorted according to the optimal expression tree to obtain a customer importance list. This customer importance list is then used in a genetic algorithm to determine the route planning for each new energy commercial vehicle. After each new energy commercial vehicle completes a node, the customer nodes in the route map are updated, and the route for each new energy commercial vehicle is dynamically planned based on the update results. By acquiring a route map containing customer, charging station, and warehouse nodes and constructing distance, energy consumption, and time matrices, the fitness of the expression tree is calculated based on the distance, energy consumption, and time matrices. Then, combined with a genetic algorithm, the initial route is determined. This fully considers key factors such as distance, energy consumption, and time. Furthermore, route planning is performed according to the customer importance list, ensuring that the needs of important customers are prioritized. Customer nodes are updated and routes are dynamically adjusted, thereby improving the operating efficiency of new energy commercial vehicles. Attached Figure Description

[0016] The invention will now be further described with reference to the accompanying drawings.

[0017] Figure 1 A flowchart illustrating an AI-based dynamic path optimization method for new energy commercial vehicles, provided as an embodiment of the present invention; Figure 2 This is a framework diagram of an AI-based dynamic path optimization system for new energy commercial vehicles, provided for an embodiment of the present invention. Detailed Implementation

[0018] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0019] Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0020] This invention provides an AI-based dynamic route optimization method for new energy commercial vehicles. See also... Figure 1 , Figure 1 A flowchart illustrating an AI-based dynamic path optimization method for new energy commercial vehicles, provided as an embodiment of the present invention. The method includes the following steps: S101, Obtain the route map of the target area and mark all valid nodes in the route map to obtain the set of valid nodes; S102, determine the driving paths between each pair of valid nodes in the valid node set to obtain the distance matrix; S103, the energy consumption matrix is ​​obtained based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and the time matrix is ​​obtained based on the distance matrix and the driving speed of the new energy commercial vehicle; S104, initialize the set of valid nodes and the set of function nodes, and randomly combine valid nodes and function nodes to construct multiple expression trees; S105: Calculate the fitness of each expression tree based on the distance matrix, energy consumption matrix, and time matrix. Sort all expression trees by fitness from largest to smallest and obtain the first preset number of expression trees for updating and optimization to obtain the optimal expression tree. S106. Sort all customer nodes according to the optimal expression tree to obtain a customer importance list, and substitute the customer importance list into the genetic algorithm to determine the route planning for each new energy commercial vehicle. S107: After each new energy commercial vehicle completes a node, the customer node on the route map is updated, and the route of each new energy commercial vehicle is dynamically planned based on the update results. The effective nodes include customer nodes, charging station nodes, and warehouse nodes; the warehouse nodes include multiple new energy commercial vehicles.

[0021] The present invention provides an AI-based dynamic route optimization method for new energy commercial vehicles. By acquiring a route map containing customer, charging station, and warehouse nodes and constructing distance, energy consumption, and time matrices, the fitness of the expression tree is calculated based on the distance, energy consumption, and time matrices. Then, a genetic algorithm is used to determine the initial route. This method fully considers key factors such as distance, energy consumption, and time. Furthermore, route planning is performed according to a customer importance list, ensuring that the needs of important customers are prioritized. Customer nodes are updated and routes are dynamically adjusted to improve the operating efficiency of new energy commercial vehicles.

[0022] In one implementation, the target area is the region where the new energy commercial vehicle is planned; the distance matrix obtained by determining the driving paths between each pair of valid nodes in the valid node set is used to import the map corresponding to the target area to determine the driving distance between each pair of valid nodes; the valid node set is initialized for customer nodes to determine the demand, energy consumption to the current vehicle, time window (preparation time / due date), and service time; vehicle status: remaining load, current battery level, and estimated arrival time; each vehicle leaves the warehouse by default with a full charge, and each customer point only needs to be allocated one new energy commercial vehicle by default; the function node set (operation rules) is initialized: binary operations: +, -, ×, ÷, max, min; unary operations: abs (absolute value), log (logarithm), neg (negation), etc.

[0023] In one implementation, a comprehensive and accurate data foundation for route planning is provided by acquiring a route map containing customer, charging station, and warehouse nodes and constructing distance, energy consumption, and time matrices. Based on these matrices, the fitness of the expression tree is calculated, and then combined with a genetic algorithm to determine the initial route. This approach fully considers key factors such as distance, energy consumption, and time, making the planned route more aligned with actual operational needs and avoiding the bias and subjectivity of traditional experience-based planning.

[0024] In one implementation, the initial preset number is determined by technical personnel.

[0025] In one implementation, route planning based on a customer importance list ensures that the needs of key customers are prioritized, improving customer satisfaction. Simultaneously, rationally arranging the driving routes of new energy commercial vehicles reduces unnecessary mileage and lowers energy costs. Furthermore, through expression tree optimization and the application of genetic algorithms, efficient utilization of vehicle resources can be achieved, avoiding vehicle idleness or overuse, thereby reducing operating costs such as vehicle maintenance.

[0026] In one implementation, after a new energy commercial vehicle completes a certain node, it updates the customer node and dynamically adjusts the route. This allows it to respond promptly to changes in customer demand (such as new orders, order cancellations, etc.), changes in traffic conditions, and changes in the vehicle's own status (such as insufficient remaining battery power). The dynamic planning capability ensures that the vehicle always travels on the optimal path, reducing delays and losses, and guaranteeing the stability and continuity of the entire operation process.

[0027] In one implementation, the combination of scientific initial route planning and subsequent dynamic adjustments can minimize vehicle travel time and improve transportation efficiency. By rationally scheduling visits to charging station nodes, vehicles can avoid being forced to stop due to insufficient battery power, ensuring continuous vehicle operation. Simultaneously, the optimized route reduces round-trip travel and waiting time, enabling vehicles to complete more transport tasks within the same timeframe, thus improving overall operational efficiency.

[0028] In one embodiment, obtaining the energy consumption matrix based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and obtaining the time matrix based on the distance matrix and the driving speed of the new energy commercial vehicle, includes: The energy consumption rate of new energy commercial vehicles is determined based on historical data, and the energy consumption matrix is ​​obtained by multiplying the energy consumption rate by the distance matrix. The average driving speed of new energy commercial vehicles is determined based on historical data, and the time matrix is ​​obtained by multiplying the average driving speed by the distance matrix.

[0029] In one implementation, the average energy consumption rate of new energy commercial vehicles is calculated based on historical driving data (energy consumption records under different road sections, loads, weather conditions, and vehicle speeds). This can truly reflect the energy consumption patterns of vehicles in actual operating scenarios. Compared with theoretical energy consumption values, the energy consumption rate derived from historical data is more in line with the complex and ever-changing actual road conditions (such as high energy consumption on uphill sections and low energy consumption on flat sections), making the subsequent calculation of the energy consumption matrix more accurate.

[0030] One implementation involves statistically analyzing historical data to determine the average travel speed between different nodes, incorporating the influence of actual factors such as traffic congestion periods, road types, and traffic light density. This is more realistic than assuming a fixed speed, ensuring that the time matrix accurately reflects the actual travel time between nodes.

[0031] In one implementation, each element (i,j) in the energy consumption matrix represents the energy consumption from effective node i to node j. This energy consumption rate is calculated by multiplying the distance (i,j) in the distance matrix by the energy consumption rate, thus transforming the physical distance between nodes into a quantifiable energy cost and providing data support for the low-energy consumption target in subsequent route planning.

[0032] In one implementation, each element (i,j) in the time matrix represents the travel time from the effective node i to the node j, which is calculated by dividing the distance (i,j) in the distance matrix by the average travel speed. This logic clearly reflects the impact of distance and speed on the time consumption, making the quantification of the time dimension more intuitive and providing an accurate basis for short-time targets.

[0033] In one embodiment, calculating the fitness of each expression tree based on the distance matrix, energy consumption matrix, and time matrix includes: Through formula Obtain the fitness F of each expression tree; Where F is the fitness of the expression tree. , and These are weighted coefficients, and their sum is 1. The total energy consumption of the expression tree. The total cost of the expression tree. The total time for the expression tree. Let be the energy consumption from node i to node j, and n be the number of nodes in the path in the expression tree. The variable is a decision variable; it is 1 if the path contains an edge from i to j, and 0 otherwise. Cost per unit of electricity Charging cost coefficient, Let $\mathbf$ be the amount of charge from node $i$ to node $j$. Here, P represents the time window penalty coefficient. Let be the travel time from node i to node j, and m be the number of client nodes in the expression tree. The service time for the kth customer.

[0034] In one implementation, If node j is a charging station, it is recorded as the charging amount; otherwise, it is 0. P is the time window penalty, and the specific value is determined by the technical personnel.

[0035] In one embodiment, sorting all customer nodes according to the optimal expression tree to obtain a customer importance list, and combining this with a genetic algorithm to determine the route planning for each new energy commercial vehicle includes: Multiple initial route plans are generated based on the customer importance list, resulting in multiple sets of parallel chromosomes; Parallel genetic operations are performed iteratively on multiple sets of parallel chromosomes until the stopping condition is met to obtain the target parallel chromosome. The route planning for each new energy commercial vehicle is determined based on the target parallel chromosome.

[0036] In one implementation, multiple initial route plans (i.e., multiple parallel chromosomes) are generated based on a customer importance list, meaning that more potential route combinations are covered from the initial planning stage. Compared to a single initial plan, this "multi-starting point" mode avoids getting stuck in local optima due to the limitations of the initial solution, finding only a relatively good route in a certain area while ignoring the globally better solution, thus providing a broader search space for subsequent optimization and improving the global optimality of the final route planning.

[0037] In one implementation, parallel genetic operations (such as parallel selection, crossover, and mutation) can process multiple sets of chromosomes simultaneously, making full use of computing resources to achieve multi-threaded optimization. Parallel processing can significantly reduce the time cost of iterative optimization, especially in scenarios with a large number of nodes and complex route combinations, it can quickly converge to a better solution, avoiding the problem of excessive time consumption in traditional serial genetic algorithms and improving the timeliness of route planning.

[0038] In one implementation, the initial route plan is generated based on a customer importance list, ensuring that the needs of important customers are prioritized from the initial planning stage. During iterative optimization, the parallel genetic operation uses customer importance as a potential constraint, prioritizing the retention of route genes that meet the needs of high-priority customers. This ensures that the final route plan achieves both global optimization and precise matching of the enterprise's customer segmentation service strategy, thereby improving the satisfaction of important customers.

[0039] In one embodiment, after each new energy commercial vehicle completes a node, the customer node on the route map is updated, and dynamic route planning for each new energy commercial vehicle is performed based on the update result, including: Update the customer nodes in the roadmap to obtain updated customer nodes; updates include removing customer nodes that have completed tasks and adding new customer nodes; Regenerate the updated optimal expression tree based on the updated client nodes, and calculate the similarity between the updated optimal expression tree and the optimal expression tree; If the similarity is greater than or equal to the preset value, the route of each new energy commercial vehicle remains unchanged; If the similarity is less than the preset value, the route for each new energy commercial vehicle will be re-determined based on the updated optimal expression tree and the genetic algorithm.

[0040] In one implementation, after a customer node is updated, the magnitude of the change can be determined by calculating the similarity between the updated optimal expression tree and the original optimal expression tree. If the similarity is greater than or equal to a preset value, it indicates that the overall change in customer demand is small, and keeping the original route unchanged can avoid problems such as vehicle detours and driver confusion caused by frequent adjustments, reducing the additional energy consumption and time costs caused by route changes. If the similarity is less than the preset value, replanning is initiated to ensure timely response when demand changes significantly, achieving a balance between minimizing disruptions to small changes and precisely adjusting for large changes. An adjacency matrix is ​​constructed from the updated optimal expression tree and the original optimal expression tree to obtain the updated adjacency matrix and the original optimal adjacency matrix. The cosine similarity between the updated adjacency matrix and the original optimal adjacency matrix is ​​calculated to obtain the similarity. The preset value is determined by technical personnel.

[0041] Based on the same inventive concept, this invention also provides an AI-based dynamic route optimization system for new energy commercial vehicles. See also Figure 2 , Figure 2 A framework diagram of an AI-based dynamic path optimization system for new energy commercial vehicles provided in this embodiment of the invention includes: The effective node set determination module is used to obtain the route map of the target area and mark all effective nodes in the route map to obtain the effective node set; the effective nodes include customer nodes, charging station nodes and warehouse nodes; the warehouse nodes include multiple new energy commercial vehicles; The first matrix determination module is used to determine the driving path between each pair of valid nodes in the valid node set to obtain the distance matrix; The second matrix determination module is used to obtain the energy consumption matrix based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and to obtain the time matrix based on the distance matrix and the driving speed of the new energy commercial vehicle. The expression tree building module is used to initialize the set of valid nodes and the set of function nodes, and to randomly combine valid nodes and function nodes to build multiple expression trees. The optimal expression tree determination module is used to calculate the fitness of each expression tree based on the distance matrix, energy consumption matrix, and time matrix, sort all expression trees from largest to smallest fitness, and obtain the top preset number of expression trees for updating and optimization to obtain the optimal expression tree; The route planning module is used to sort all customer nodes according to the optimal expression tree to obtain a customer importance list, and then substitute the customer importance list into the genetic algorithm to determine the route planning for each new energy commercial vehicle. The route update module is used to update the customer nodes on the route map after each new energy commercial vehicle has completed a node, and to dynamically plan the route for each new energy commercial vehicle based on the update results.

[0042] The AI-based dynamic route optimization system for new energy commercial vehicles provided in this invention obtains a route map containing customer, charging station, and warehouse nodes and constructs distance, energy consumption, and time matrices. Based on these matrices, the fitness of the expression tree is calculated, and a genetic algorithm is used to determine the initial route. This system fully considers key factors such as distance, energy consumption, and time. Furthermore, route planning is performed according to a customer importance list, ensuring that the needs of important customers are prioritized. The system updates customer nodes and dynamically adjusts routes, thereby improving the operating efficiency of new energy commercial vehicles.

[0043] In one embodiment, the second matrix determination module includes: The energy consumption matrix determination module is used to determine the energy consumption rate of new energy commercial vehicles based on historical data, and obtain the energy consumption matrix by multiplying the energy consumption rate by the distance matrix. The time matrix determination module is used to determine the average driving speed of new energy commercial vehicles based on historical data, and obtain the time matrix by multiplying the average driving speed by the distance matrix.

[0044] In one embodiment, the optimal expression tree determination module includes: Through formula Obtain the fitness F of each expression tree; Where F is the fitness of the expression tree. , and These are weighted coefficients, and their sum is 1. The total energy consumption of the expression tree. The total cost of the expression tree. The total time for the expression tree. Let be the energy consumption from node i to node j, and n be the number of nodes in the path in the expression tree. The variable is a decision variable; it is 1 if the path contains an edge from i to j, and 0 otherwise. Cost per unit of electricity Charging cost coefficient, Let $\mathbf$ be the amount of charge from node $i$ to node $j$. Here, P represents the time window penalty coefficient. Let be the travel time from node i to node j, and m be the number of client nodes in the expression tree. The service time for the kth customer.

[0045] In one embodiment, the route planning module includes: A multi-parallel chromosome generation module is used to generate multiple initial route schemes based on a customer importance list, resulting in multiple sets of parallel chromosomes. The target parallel chromosome determination module is used to iteratively perform parallel genetic operations on multiple sets of parallel chromosomes until the stopping condition is met to obtain the target parallel chromosome. The new energy commercial vehicle route determination module is used to determine the route planning for each new energy commercial vehicle based on the target parallel chromosome.

[0046] In one embodiment, the route update module includes: The customer node update module is used to update customer nodes in the roadmap to obtain updated customer nodes; the update includes removing customer nodes that have completed tasks in the roadmap and adding new customer nodes; The similarity calculation module is used to regenerate the updated optimal expression tree based on the updated client node and calculate the similarity between the updated optimal expression tree and the optimal expression tree. The route-unchanged module is used to ensure that the route of each new energy commercial vehicle remains unchanged if the similarity is greater than or equal to a preset value. The route change module is used to redetermine the route of each new energy commercial vehicle based on the updated optimal expression tree and combined with a genetic algorithm if the similarity is less than a preset value.

[0047] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A dynamic path optimization method for new energy commercial vehicles based on AI, characterized in that, The method includes: Obtain a route map of the target area and mark all valid nodes in the route map to obtain a set of valid nodes; the valid nodes include customer nodes, charging station nodes, and warehouse nodes; the warehouse nodes include multiple new energy commercial vehicles; The distance matrix is ​​obtained by determining the driving paths between each pair of valid nodes in the set of valid nodes. An energy consumption matrix is ​​obtained based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and a time matrix is ​​obtained based on the distance matrix and the driving speed of the new energy commercial vehicle. Initialize the set of valid nodes and the set of function nodes, and randomly combine the valid nodes and function nodes to construct multiple expression trees; The fitness of each expression tree is calculated based on the distance matrix, energy consumption matrix, and time matrix. All expression trees are sorted from largest to smallest fitness, and the top preset number of expression trees are obtained for updating and optimization to obtain the optimal expression tree. The customer importance list is obtained by sorting all customer nodes according to the optimal expression tree, and the customer importance list is substituted into the genetic algorithm to determine the route planning for each new energy commercial vehicle. After each new energy commercial vehicle completes a node, the customer node in the route map is updated, and the route of each new energy commercial vehicle is dynamically planned based on the update result.

2. The AI-based dynamic path optimization method for new energy commercial vehicles according to claim 1, characterized in that, An energy consumption matrix is ​​obtained based on the distance matrix and the energy consumption rate of the new energy commercial vehicle. A time matrix is ​​obtained based on the distance matrix and the driving speed of the new energy commercial vehicle, including: The energy consumption rate of the new energy commercial vehicle is determined based on historical data, and the energy consumption matrix is ​​obtained by multiplying the energy consumption rate by the distance matrix. The average driving speed of the new energy commercial vehicle is determined based on historical data, and the time matrix is ​​obtained by multiplying the average driving speed by the distance matrix.

3. The AI-based dynamic path optimization method for new energy commercial vehicles according to claim 1, characterized in that, The fitness of each expression tree is calculated based on the distance matrix, energy consumption matrix, and time matrix, including: Through formula Obtain the fitness F of each expression tree; Where F is the fitness of the expression tree. , and These are weighted coefficients, and their sum is 1. The total energy consumption of the expression tree. The total cost of the expression tree. The total time for the expression tree. Let be the energy consumption from node i to node j, and n be the number of nodes in the path in the expression tree. The variable is a decision variable; it is 1 if the path contains an edge from i to j, and 0 otherwise. Cost per unit of electricity Charging cost coefficient, Let $\mathbf$ be the amount of charge from node $i$ to node $j$. Here, P represents the time window penalty coefficient. Let be the travel time from node i to node j, and m be the number of client nodes in the expression tree. The service time for the kth customer.

4. The AI-based dynamic path optimization method for new energy commercial vehicles according to claim 1, characterized in that, Based on the optimal expression tree, all customer nodes are sorted to obtain a customer importance list. Then, a genetic algorithm is used to determine the route planning for each new energy commercial vehicle, including: Multiple initial route schemes are generated based on the customer importance list, resulting in multiple sets of parallel chromosomes; Parallel genetic operations are performed iteratively on multiple sets of parallel chromosomes until the stopping condition is met to obtain the target parallel chromosome. The route planning for each new energy commercial vehicle is determined based on the target parallel chromosome.

5. The AI-based dynamic path optimization method for new energy commercial vehicles according to claim 4, characterized in that, After each new energy commercial vehicle completes a node, the customer node in the route map is updated. Based on the update result, dynamic route planning is performed for each new energy commercial vehicle, including: The customer nodes in the roadmap are updated to obtain updated customer nodes; the update includes removing customer nodes in the roadmap that have completed their tasks and adding new customer nodes. Based on the updated client node, regenerate the updated optimal expression tree and calculate the similarity between the updated optimal expression tree and the optimal expression tree; If the similarity is greater than or equal to a preset value, the route of each new energy commercial vehicle remains unchanged. If the similarity is less than a preset value, the route of each new energy commercial vehicle is re-determined based on the updated optimal expression tree and the genetic algorithm.

6. An AI-based dynamic path optimization system for new energy commercial vehicles, characterized in that, The system includes: The effective node set determination module is used to obtain a route map of the target area and mark all effective nodes in the route map to obtain an effective node set; the effective nodes include customer nodes, charging station nodes and warehouse nodes; the warehouse nodes include multiple new energy commercial vehicles; The first matrix determination module is used to determine the driving path between each pair of valid nodes in the set of valid nodes to obtain a distance matrix; The second matrix determination module is used to obtain an energy consumption matrix based on the distance matrix and the energy consumption rate of the new energy commercial vehicle, and to obtain a time matrix based on the distance matrix and the driving speed of the new energy commercial vehicle. An expression tree construction module is used to initialize the set of valid nodes and the set of function nodes, and to randomly combine valid nodes and function nodes to construct multiple expression trees; The optimal expression tree determination module is used to calculate the fitness of each expression tree based on the distance matrix, energy consumption matrix, and time matrix, sort all expression trees from largest to smallest fitness, and obtain the top preset number of expression trees for updating and optimization to obtain the optimal expression tree; The route planning module is used to sort all customer nodes according to the optimal expression tree to obtain a customer importance list, and then substitute the customer importance list into the genetic algorithm to determine the route planning for each new energy commercial vehicle. The route update module is used to update the customer nodes in the route map after each new energy commercial vehicle has completed a node, and to dynamically plan the route for each new energy commercial vehicle based on the update results.

7. The AI-based dynamic path optimization system for new energy commercial vehicles according to claim 6, characterized in that, The second matrix determination module includes: The energy consumption matrix determination module is used to determine the energy consumption rate of the new energy commercial vehicle based on historical data, and obtain the energy consumption matrix by multiplying the energy consumption rate by the distance matrix. The time matrix determination module is used to determine the average driving speed of the new energy commercial vehicle based on historical data, and to obtain the time matrix by multiplying the average driving speed by the distance matrix.

8. The AI-based dynamic path optimization system for new energy commercial vehicles according to claim 6, characterized in that, The optimal expression tree determination module includes: Through formula Obtain the fitness F of each expression tree; Where F is the fitness of the expression tree. , and These are weighted coefficients, and their sum is 1. The total energy consumption of the expression tree. The total cost of the expression tree. The total time for the expression tree. Let be the energy consumption from node i to node j, and n be the number of nodes in the path in the expression tree. The variable is a decision variable; it is 1 if the path contains an edge from i to j, and 0 otherwise. Cost per unit of electricity Charging cost coefficient, Let $\mathbf$ be the amount of charge from node $i$ to node $j$. Here, P represents the time window penalty coefficient. Let be the travel time from node i to node j, and m be the number of client nodes in the expression tree. The service time for the kth customer.

9. The AI-based dynamic path optimization system for new energy commercial vehicles according to claim 6, characterized in that, The route planning module includes: A multi-parallel chromosome generation module is used to generate multiple initial route schemes based on the customer importance list to obtain multiple sets of parallel chromosomes. The target parallel chromosome determination module is used to iteratively perform parallel genetic operations on multiple sets of parallel chromosomes until the stopping condition is met to obtain the target parallel chromosome. The new energy commercial vehicle route determination module is used to determine the route plan for each new energy commercial vehicle based on the target parallel chromosome.

10. The AI-based dynamic path optimization system for new energy commercial vehicles according to claim 9, characterized in that, The route update module includes: The customer node update module is used to update the customer nodes in the roadmap to obtain updated customer nodes; the update includes removing customer nodes in the roadmap that have completed tasks and adding new customer nodes; The similarity calculation module is used to regenerate the updated optimal expression tree based on the updated client node, and calculate the similarity between the updated optimal expression tree and the optimal expression tree. The route-unchanged module is used to ensure that the route of each new energy commercial vehicle remains unchanged if the similarity is greater than or equal to a preset value. The route change module is used to redetermine the route of each new energy commercial vehicle based on the updated optimal expression tree and the genetic algorithm if the similarity is less than a preset value.

Citation Information

Patent Citations

  • Vehicle distribution path optimization method based on genetic algorithm

    CN117132010A

  • Unmanned ship navigation route planning method and system

    CN118795896A

  • Vehicle route optimization method based on K-Medoids clustering and improved genetic algorithm

    CN120046827A

Cited By

  • Vehicle scheduling optimization method and system based on real-time position and energy consumption

    CN121961377A